| Summary: | 博士 === 國立中山大學 === 電機工程學系研究所 === 106 === This dissertation is focused on the improvement of state-of-charge (SOC) estimation with the coulomb counting method for rechargeable batteries by means of state-of-health (SOH) calibration. The proposed approach intends to provide an easy-to-use solution for on-line indication with high accuracy to estimate the battery status without the need of demanding calculations or hard-earned databases. To estimate the SOC of an aged battery more accurately, the degradation of its full capacity has to be taken into account. By scheduling the battery’s charging/discharging current and monitoring the battery’s status, the existing full capacity can be updated regularly by normal calibration or occasionally by partial calibration, in which the charging/discharging rates are normalized with the latest updated full capacity to agree with the battery’s statuses. To exclude the estimation error caused by an inaccurate current, the SOC is reset to 0% when the battery is completely exhausted and 100% for a fully charged battery. With an updated SOH, the battery C-rate is re-scaled accordingly. Experimental tests are carried out to demonstrate that the proposed method can provide an accurate online indication of batteries’ SOCs. With an implanted error of 0.3% in current measurement, the SOC estimation error can always be less than 1.905% after a number of SOH calibrations.
On the other hand, a laboratory battery power system is set up by four buck-boost type battery power modules (BPMs) with the lithium iron phosphate (LiFePO4) batteries to verify the proposed SOC estimation method with SOH calibration. During the operation of the battery power bank, the BPMs are controlled individually and also work collaboratively to meet the load requirements or sharing the charging power from the DC source. Charge equalization among batteries can be executed by individually controlling the currents either into or out from the batteries in accordance with the real-time SOCs, estimated with the proposed estimation. By identifying the maximum charge status with normal or partial calibration, the SOHs of the batteries in BPMs can be updated accordingly, and then the SOCs can be estimated more precisely. Moreover, without interrupting the system operation, the exhausted or damaged battery can be isolated simply by turning off its corresponding active power switch of the BPM without the need of an extra mechanical switch. Experimental tests are carried out to demonstrate that the battery power system can provide an accurate online indication of batteries’ SOCs with properly programmed discharging/charging scenarios to conduct charge equalization, output voltage regulation, and fault tolerance throughout discharging/charging processes.
|
|---|
